A Robust Approach to Facies Inversion using Relative Rock Physics

Summary This study investigates the use of relative rock physics, specifically cross-plotting relative Lambda vs. relative Mu, to distinguish different facies and predict them away from the well location. Relative properties are computed by removing the low frequency trend from the absolute rock properties. A rock physics template was created using well log data and shale as the background. Sands were modeled using Gassmann fluid substitution and a calibrated pore space stiffness model. Seismic-derived relative Lambda and Mu were estimated using a linearized AVO equation and then phase-rotated to produce relative properties. The resulting cross-plot shows how different facies such as carbonates, source rocks, and hydrocarbon or water-bearing sands cluster in relative property space. By identifying boundaries between facies in this space, the method enables facies volumes to be generated and extended beyond the well. The final inversion successfully identifies water and hydrocarbon saturated prospects, demonstrating that facies can be predicted without needing the low frequency models or absolute rock properties, unlike traditional facies inversion approaches.